As I’m writing this, there seem to be a huge amount of furor over the US elections. I don’t want to get into that particular mess, but I had a few idle thoughts about how to fix one of the root issues here. The fact that the system of voting we use is based on pen & paper and was designed when a computer was a job title for a practical mathematician.
A large part of the problem is that we have to wait for a particular date, and there are polls, which seems to be used for information and misinformation at the same time. This means that people are basically trying to vote with insufficient information. It would be much more straightforward if the whole system was transparent and iterative.
Note: This is a blog post I’m writing to get this idea out of my head. I did very little research and I’m aware that there is probably a wide body of proposals in the area, which I didn’t look at.
The typical manner I have seen suggested for solving the election problem is to ensure the following properties:
- Verifying that my vote was counted properly.
- Verifying that the total votes were counted fairly.
- (maybe) Making it hard / impossible to show a 3rd party who I voted to.
The last one is obviously pretty hard to do, but is important to prevent issues with pressuring people to vote for particular candidates.
I don’t have the cryptographic chops to come up with such a system, not do I think that it would be particularly advisable. I would rather come up with a different approach entirely. In my eyes, we can have a system where we have the following:
- The government issues each voter with a token (coin) that they can spend. That will probably get you to think about blockchains, but I don’t think this is a good idea. If we are talking about technical details, let’s say that the government issues a certificate to each voter (who generates their own private key, obviously).
- The voter can then give that voting coin to a 3rd party. For example, but signing a vote for a particular candidate using their certificate.
- These coins can then be “spent” during election. The winner of the election is the candidate that got more than 50% of the total coins spent.
As it currently stands, this is a pretty horrible system. To start with, this means that it is painfully obvious who voted for whom. I think that a transparent voting record might be a good idea in general, but there are multitude of problems with that. So like many great ideas in theory, we shouldn’t allow it.
This is usually where [complex cryptography] comes into play, but I think that a much better option would be to introduce the notion of brokers into the mix. What do I mean?
While you could spend your coin directly on a particular candidate, you could also spend it at a broker. That broker is then delegated to use your vote. You may use a party affiliated broker or a neutral one. Imagine a 2 party system when you have the Olive party and the Opal party. I’m using obscure colors here to try to reduce any meaning people will read into the color choices. For what it’s worth, red & blue as party colors have opposite meaning in the states and Israel, which is confusing.
Let’s take two scenarios into considerations:
- A voter spend their coin on the Olive Political Action Committee, who is known to support Olive candidates. In this case, you can clearly track who they want to vote for. Note that they aren’t voting directly for a candidate, because they want to delegate their vote to a trusted party to manage that.
- A voter spend their coin on a Private Broker #435. While they do that, they instruct the broker to pass their vote to the Olive PAC broker, or a particular candidate, etc.
The idea is that the private broker is taking votes from enough people that while it is possible to know that you went through a particular broker, you can’t know who you voted for. The broker itself obviously know, but that is similar to tracking the behavior of a voting booth, which also allows you to discover who voted to whom. I think that it is possible to set things up so the broker itself won’t be able to make this association, however. Secured Sum would work, probably. A key point for me is that this is an issue that is internal to a particular broker, not relevant to the system as a whole.
So far, I don’t think that I’m doing too much, but the idea here is that I want to take things further. Instead of stopping the system there, we allow to change the vote. In other words, instead of having to vote blindly, we can look at the results and adjust them.
In the Apr 2019 Israeli election, over 330 thousands votes were discarded because they didn’t reach the minimum threshold. That was mostly because the polls were wrong, because I don’t think that people would have voted for those parties if they knew that they are on the verge. Being able to look at that and then adjust the votes would allow all those votes to be counted.
Taking this further, once we have the system of brokers and electronic votes in place, there is no reason to do an election once every N years. Instead, we can say that the coins are literal political capital. In order to remain in office, the elected officials must keep holding over 50% of the amount of spent coins. It would probably be advisable to actually count these on a weekly / bi-weekly basis, but doing this on a short time intervals means that there is going to be a lot more accountability.
Speaking from Israel’s point of view, there have been sadly numerous cases where candidates campaigned on A, then did the exact opposite once elected. There is even a couple of famous sayings about it:
- We promised, but didn’t promise to fulfil.
- What you see from here you can’t see from there.
Note that this is likely to result in more populist system, since the elected officials are going to pay attention to the electorate on an ongoing basis, rather than just around election time. I can think of a few ways to handle that. For example, once seated, for a period of time, you’ll need > 50% of the coins to get an elected official out of office.
A few more details:
- For places like the states, where you vote for multiple things at the same time (local, state, federal house & senate, president), you’ll get multiple coins to spend, and can freely spend them in different locations. Or, via a broker, designate that they are to be spend on particular directions.
- A large part of the idea is that a voter can withdraw their coin from a broker or candidate at any time.
- Losing the key isn’t a big deal. You go to the government, revoke your pervious certificate and get a new one. The final computation will simply ignore any revoked coins.
- The previous point is also important to dealing with people who died or moved. It is trivial to ensure that the dead don’t vote in this scheme, and you can verify that a single person don’t have coins from multiple different locations.
The implications of such a system are interesting, in my eyes. The idea is that you delegate the vote to someone you trust, directly or indirectly. That can be a candidate, but most likely will be a group. The usual term in the states in PAC, I believe. The point is that you then have active oversight by the electorate on the elected officials.
Over seven years ago I wrote a post about what I most dislike in the current election system. You are forced to vote on your top issue, but you usually have a far more complex system of values that you have to balance. For example, let’s say that my priorities are:
- National security
- Fiscal policy
- Access to violet flowers
I may have to balance between a candidate that want to ban violet flowers but propose to lower taxes or a candidate that wants to raise taxes and want to legalize violet flowers. Which one do I choice? If I can shift my vote as needed, I can make that determination at the right political time. During the budget month, my votes goes to $PreferredFiscalPolicy candidate and then if they don’t follow my preferences on violet flowers, I can shift.
This will have the benefit of killing the polls industry, since you can just look at where the political capital is. And it will allow the electorate to have a far greater control over the government. I assume that elected officials will then be paying very careful attention to how much political capital they have to spend and act accordingly.
I wonder if I should file this post under science fiction, because that might make a good background for world building. What do you think of this system? And what do you think the effects of it would be?